RAM overclocked works fine. 1 memtest error.

[drivinfast247]

I recently upgraded my system from a R5 1600x and 2666mhz RAM(overclocked to 2933mhz) to a R5 2600x and G.Skill Flare X 3200mhz(overclocked to 3600mhz and cl16). I used the same Asrock Fatality B350 micro-itx board and Corsair CX650m PSU(I know, I know, new one is on order). And EVGA GTX 1080 SC GPU.

The computer runs like a charm and is a beast at games 144hz w/Gsync. For the heck of it a decided to run memtest while I went to work. I came home early to check on it and it ran for over 7 hours and was still running. Buuuuuut.....it did show one error.

Is one error too many? What/why the error?

 

[jossrik]

There are many many reasons for mem errors, with overclocking RAM it's gonna happen. No way around it. That being said, there are different types of stable for those of us who overclock. I have a 32Gb kit on a 2700X, x470 Taichi with stupid fast timings, I can start my computer, and game for hours. No problems. Soon as I run a stress ap though it fails almost immediately. For me, it's fine. I do see people who say memtest for 24 hours no errors is stable. I think that's a bit extreme though. 7 hours one error could be a stray bit of radiation that made it through the atmosphere. Might point to instability, might not is what I'm saying, there are tons of reasons for it to have errors, not enough voltage, too much voltage, the list goes on and on.

 

[clutchc]

Probably due to the RAM OC. If you want the system to be 100% stable, back off on the RAM OC. If you're happy as-is, live with the one error. But to answer your question, yes. One error usually represents a defect in the RAM system.

 

[drivinfast247]

Appreciate the answers. If I start getting Windows errors or in game errors or crashing I'll back it down. I'm surprised I got it this stable with those components. Ran Prime95 Blend for 15 mins without a hitch.

 

[drivinfast247]

Ok, I'm back. I went to Best Buy and picked up a new PSU and 1tb SSD. The PSU was a Corsair RM 850x. The SSD is a 860 EVO. I installed them in the case and moved my Steam library to the SSD.

Yesterday I was playing a few games and then played ROTTR. The game played fine and temps were always in check. GPU 70c or under and CPU 55c or less. I decided to run the ingame benchmark and I had a BSOD code Kernalpower 41. Thought that was odd. I went ahead and lowered the RAM OC to 3466mhz and used the timings suggested by the ryzen ram app. Played a few more hours of random games ie:Arkham Knight, Metro LL redux and Dying light. No issues. Yesterday and today I had two different games just exit to desktop. Neither are showing on the event viewer. Is there a way to recover a crash dump without it showing in Event viewer?

I know that PSU is not the best, but Best Buy really only carries that brand. They do carry the HX850i model. Maybe I'll swap out PSUs.

I am going to run the ram at stock speeds tomorrow and run a few passes of memtest86 before I assume a PSU issue.
tl;dr
So this long drawn on post was to ask if there is a way to recover a crash dump not showing in event viewer? And can a ram overclock cause games to quit to desktop? Also should I even bother with the PSU?

 

[Pavkata882]

I think it shouldnt bother you. My system is running with 4 different ram sticks but have no crashes whatsoever. Though, my system also cant pass memtest with less than 1 error, but even then its perfectly stable.

 

[Gaidax]

I don't understand how this thing is being spread and endorsed - memory errors are BAD.

Do you even understand what this means? What if tomorrow you will have this error affect system files and result you in having to bloody reinstall everything from scratch? Because that's what can happen. You are literally playing with fire there - you will be gathering errors, dumping them on disc and eventually you will have things crashing right and left several months down the road without ability to fix it without nuking all down, because you already did permanent damage to files and apps.

Biggest problem with such errors is not the apparent stuff - error causing crashing is your best case, worst case is when things appear working while in fact they don't and in long haul this results in overall system instability simply because of all the errors you hoarded, wrong shit written and what not.

There is a reason why smart people say that stable OC = no errors with good thorough testing, ESPECIALLY for RAM.

 

[Pavkata882]

You may be right but it has been running like that for over an year with the settings. At the beginning i was, of course, worried and did daily backups of imoprant files, but after a while i stopped caring since ive read many times that a lot of people have no issues but yet the have some errors in memtest. And also it may be the reason that the system is operating 4 different sticks. Not arguing with your opinion though, dont take it too seriously, we are all here to share experience overall. Cheers!

 

[drivinfast247]

Update: I ran memtest for 6 hours and it made 3 clean passes, no errors. The RAM was clocked at 3466mhz.

And also still wondering if I can find/produce a dump file after a game quits to desktop?

 

[clutchc]

Update: I ran memtest for 6 hours and it made 3 clean passes, no errors. The RAM was clocked at 3466mhz.

And also still wondering if I can find/produce a dump file after a game quits to desktop?

Then I wouldn't worry about it any more.

 

[drivinfast247]

Update: I was still getting very random games the would crash to desktop. Even though the RAM made 3 clean passes, I think it was still giving me issues. I set the ram to stock 3200 and used fast timings and it's been great! No crashes in 4 days.